Translation systems typically provide the most likely translation result to the user even though the “most likely translation result” may be incorrect. This can be the case for desktop as well as mobile devices. Commonly, if a user wishes to obtain a new translation, such as to replace an incorrect translation, the user must start over and hope the translation engine provides a different result. This can be true for both text input devices and mobile devices. For example, some devices that provide speech translation capabilities require restarting when an incorrect translation result is output. The methods for restarting may include re-entering text in a desktop application, or turning a device off and on, or providing some other input or signal to the device that it is to restart the translation. It would be beneficial if a device could provide a simple user interface that provided an alternate translation result without restarting the translation process, repeating inputs or requiring complex input sequences into the device.